Comparison of the Tests Used to Identify or Screen for Frailty

Comparison of the Tests Used to Identify or Screen for Frailty A. 18 High-Yield Points The Fundamental Divide

All frailty tools fall into two conceptual families, and this is the single most important contrast to state: - Phenotype model — frailty is a biological syndrome with a pathophysiological substrate, expressed as a limited number of physical domains. Task and performance based. - Cumulative deficit model — frailty is the accumulation of deficits (diseases, impairments, abnormal findings) and reflects biological ageing. Count based.

Tools by family: - Phenotype: Fried frailty phenotype, Frailty Trait Scale, FRAIL index - Cumulative deficits:

A second, equally important divide is purpose: screening or case-finding (rapid, high sensitivity, applied to everyone) versus assessment or diagnosis (detailed, drives the care plan). Screening is never an end in itself — a positive screen must trigger a comprehensive geriatric assessment.

Reproduce this table verbatim: Cardiovascular Health Study (Fried) Study of Osteoporotic Fractures (SOF) IANA FRAIL scale Weight loss (10 lb in a year) Weight loss Fatigue Exhaustion (self-report) Inability to rise from a chair 5 times without using Resistance (climb one flight of arms stairs) Weakness (grip strength, lowest 20%) Reduced energy level Aerobic (walk one block) Walking speed (15 ft, slowest 20%) — Illnesses (more than 5) Low physical activity (kcal/week, lowest — Loss of weight (5%)

Contrast within the table: Fried requires equipment (dynamometer, stopwatch, activity questionnaire) and population-based cut-offs (lowest 20%); SOF and FRAIL were developed because they are easier to use in the clinic and are validated. FRAIL is purely self-report — no equipment at all.

Scoring is common to all three: 0 equals robust, 1–2 equals pre-frail, 3 or more equals frail. This shared cut-off makes them directly comparable even though the items differ.

Fried phenotype — the reference standard. Advantages: biologically grounded, identifies a treatable pre-disability state, gives the prevalence figure of about 6.9%. Disadvantages: time-consuming, needs equipment and normative data, excludes cognition and mood, cannot be applied to a bed-bound or wheelchair-dependent patient.

SOF index — the pragmatic three-item version. Advantage: no grip dynamometer or activity questionnaire; the chair-stand replaces both. Disadvantage: even narrower domain coverage than Fried.

FRAIL scale — the fastest phenotype screen. Advantage: five questions, purely verbal, usable over the telephone; the "I" item uniquely captures multimorbidity, and the tool maps directly onto a management algorithm. Disadvantage: self-reported, so subject to recall bias and denial.

Frailty Index (Rockwood–Mitnitski) — counts deficits and expresses them as a proportion (deficits present divided by deficits counted). Advantages: continuous variable, grades severity finely, predicts outcomes well, and can be built from existing records. Disadvantages: needs 30 or more variables, laborious by hand, and it is not a syndrome — it deliberately mixes disease, disability and impairment; it can be considered a comorbidity index.

Clinical Frailty Scale — a global clinical judgement on a 9-point pictorial scale. Advantages: seconds to apply, no equipment, works in acutely unwell patients. Disadvantages: subjective and rater-dependent, conflates frailty with disability at the severe end.

Frailty Trait Scale — phenotype-based but multidimensional; used mainly in research and in ADVANTAGE/ICFSR recommendations.

VES-13 (Vulnerable Elders Survey-13) — a self-report tool for identifying vulnerable older people in the community, measuring age, self-rated health, physical function and ADL disability. It has been compared directly against full CGA in older men with prostate cancer on androgen deprivation.

aCGA (abbreviated CGA) — a pre-screen in older cancer patients, with published scoring and interpretation. VES-13 and aCGA screen for vulnerability, not for the frailty phenotype — they exist to decide who needs a full CGA, not to diagnose a syndrome.

SIOG task force position: screening tools exist and have been used successfully in settings such as the emergency room, but globally they were poorly tested, and the task force could not recommend any specific tool or approach above others — general geriatric experience should guide the choice.

Principle of hierarchy: frailty testing is only worthwhile in the robust-to-pre-frail range. If a patient is already severely ADL-dependent, do not measure frailty — the ADL deficits, not the frailty score, determine risk and management.

Setting dictates the tool. Adapt the assessment to the patient's complexity and to the place where it is carried out — a brief screen in the office or home, a fuller interdisciplinary work-up in an inpatient or day-hospital setting.

Whichever tool is used, the endpoint is the same: a positive screen must be followed by a search for the causative factors (disease, pain, poor nutrition, sarcopenia, executive dysfunction) and a multifaceted treatment regimen — because frailty may have a single reversible cause such as anaemia or polymyalgia rheumatica.

Model Phenotype (syndrome, treatable) versus cumulative deficits (risk index, prognostic)

Purpose Screen (sensitive, fast) versus assess (specific, comprehensive)

Method Self-report versus performance-based versus clinician judgement versus automated record-based 4. Output Categorical (robust/pre-frail/frail) versus ordinal (CFS 1–9) versus continuous (FI 0–1)

Overlap with disability Fried and FRAIL exclude disability by design; CFS and FI include it 6. Actionability FRAIL maps to an intervention algorithm; FI and CFS mainly stratify prognosis D. Latest Updates (2023–Present)

Head-to-head diagnostic accuracy — the emergency department evidence. A systematic review and meta-analysis of prospective diagnostic accuracy studies (2000–2022) covering seven ED frailty screens (PRISMA-7, CFS, VIP, FRESH, BPQ, TRST, ISAR) in 1,663 participants found a pooled sensitivity of 0.85 and specificity of 0.77 against CGA; PRISMA-7 was the most sensitive (0.84–0.98) but only moderately specific (0.64–0.78), whereas CFS was the least sensitive but the most specific.

Direct comparison against CGA. In 197 ED attendees aged 70 and above (49% frail by CGA), the CFS (AUROC 0.91) and PRISMA-7 (AUROC 0.90) were significantly more accurate than the 3- and 4-item VIP (both 0.84), with no significant difference between CFS and PRISMA-7. In a separate cohort of 313 patients (58% frail by CGA), PRISMA-7 was most accurate (AUC 0.88), followed by CFS (0.83) and ISAR (0.78).

PRISMA-7 systematic review (2025). PRISMA-7 is completed in under 2 minutes, showed an AUC of 0.85 against the Fried phenotype, and is a practical alternative to the time-consuming and resource-intensive phenotype.

Automated screening has arrived. The electronic frailty index (eFI2) is replacing the original eFI across English GP systems, available from 2025. The recommended sequence is now eFI to identify the cohort → Clinical Frailty Scale to confirm → CGA to assess and plan. This is the key structural change: screening is no longer clinician-initiated but population-triggered.

NEJM 2024 frailty-guided care. Kim and Rockwood define robust as Fried phenotype 0, Frailty Index below 0.10, and CFS 1–3, and stratify management across four stages. The tools are now used not just to label but to decide how intensively to treat.

Guideline position on choice of tool. ICFSR: all adults aged 65 and over should be offered screening for frailty using a validated rapid frailty instrument suitable to the specific setting or context (strong recommendation, low certainty), followed by clinical assessment of all who screen positive. In primary care, the FRAIL scale can be answered in 15 to 30 seconds, predicted disability and mortality at 9 years, and performed as well as the Fried phenotype and the SOF index.

Oncology has standardised on a single tool. The German S3 guideline recommends specific screening instruments mainly in the oncological setting (the G8 questionnaire), and ASCO 2023 recommends the Practical Geriatric Assessment (PGA).

Indian data — the tool changes the answer. In a meta-analysis of 330,007 community-dwelling Indians, pooled frailty prevalence was 36%, but varied markedly by instrument — 48% by frailty index versus 31% by Fried phenotype. In LASI, the most discriminant phenotype component was grip strength in men (79.1%) and physical activity in women (81.6%), arguing for sex-specific interpretation.